ToolTrust
B17/100
qdrant

mcp-server-qdrant

mcp0.8.1

@qdrant

An official Qdrant Model Context Protocol (MCP) server implementation

By qdrant | 5 findings | Scanned 3/22/2026 | tooltrust-scanner/v0.1.15

1 High3 Medium1 Low

Security Findings (5)

  • HighExcessive Permission Surface

    search_context:tool declares network permission

    Rule: AS-002Fix: Tool requests broad permissions (exec/fs/network). Validate input parameters using Enums where possible, and restrict file system operations to explicit allowed directories.
  • MediumExcessive Permission Surface ×3
    • update_context:tool declares fs permission
    • delete_context:tool declares fs permission
    • update_stream_context:tool declares fs permission
    Rule: AS-002Fix: Tool requests broad permissions (exec/fs/network). Validate input parameters using Enums where possible, and restrict file system operations to explicit allowed directories.
  • LowDoS Resilience — Missing Rate Limit / Timeout

    search_context:tool performs network or execution operations but declares no rate-limit, timeout, or retry configuration

    Rule: AS-011Fix: Declare explicit rate-limit, timeout, and retry configuration for all network and execution tools. Implement exponential back-off and surface resource state to the calling agent.

Scan this tool yourself

Reproduce this audit locally or block risky tools in CI.

Install once, then scan any MCP server:

$ curl -sfL https://raw.githubusercontent.com/AgentSafe-AI/tooltrust-scanner/main/install.sh | bash
$ tooltrust-scanner scan --server "npx -y mcp-server-qdrant"

Adjust the package name if your npm registry name differs from the tool ID. View source

Add badge to your README

Copy this Markdown to show your ToolTrust grade on GitHub.

[![ToolTrust Grade B](https://raw.githubusercontent.com/AgentSafe-AI/tooltrust-directory/main/docs/badges/grade-b.svg)](https://github.com/AgentSafe-AI/tooltrust-directory)